Forecasting volatility in the financial markets: a comparison of alternative distributional assumptions

Abstract

This article analyses the volatility forecasting performance of the GARCH models based on various distributional assumptions in the context of stock market indices and exchange rate returns. Using rollover methods to construct the out-of-the-sample volatility forecasts, this study shows that the GARCH model combined with the logistic distribution, the scaled student's t distribution and the Riskmetrics model are preferable both in stock markets and foreign exchange markets. The exponential power and the mixture of two normal distributions are, however, less recommended. Furthermore, a complex distribution does not always outperform a simpler one, although the exact ranking depends on the application of underlying assets and the performance statistics being used.

Forecasting S&P-100 stock index volatility: The role of volatility asymmetry and distributional assumption in GARCH models

“…The results show that the GARCH-N model is not outperformed by the competing models with returns innovations that allow for fat-tailed, leptokurtic and skewed characteristics. 9 These findings are consistent with the results of Chuang et al (2007), which showed that a complex distribution does not always outperform a simpler one. Nevertheless, our results contradict earlier findings of Wilhelmsson (2006), which argued that the GARCH model with a t-distribution improves volatility forecasts.…”

“…The first category exploits symmetric GARCH models with alternative distributional assumptions to forecast volatility in various financial markets. Examples include Wilhelmsson (2006) and Chuang et al (2007). Wilhelmsson (2006) the predictive ability of the GARCH(1, 1) model when estimating S&P-500 index future returns with various error distributions.…”

“…Notably, allowing for time-varying skewness or kurtosis of the distribution does not further improve forecasts. Recently, Chuang, Lu, and Lee (2007) compared the volatility forecasting performance of linear GARCH models based on a group of distributional assumptions in the context of equity and foreign exchange markets. Using moving window estimation, they showed that a complex distribution does not always outperform a simpler one.…”

Volatility modeling for IDR exchange rate through APARCH model with student-t distribution

“…As pointed out in the previous literature (e.g., see [3,17,18]), the Student-t distribution as an assumption of distribution in GARCH model outperforms some alternative distributions such as the normal, exponential, and mixture of normal distributions. If has the Student-t distribution, the probability density functions of and are respectively given by, for t = 1, ..., T,…”

Forecasting accuracy of stochastic volatility, GARCH and EWMA models under different volatility scenarios

“…For our purposes, we will use the above 'classic' GARCH model with p ¼ q ¼ 1. The range of possible models has been investigated by several authors; for example, McMillan et al (2000) compare the forecast accuracy of a variety of GARCH formulations; Chuang et al (2007) investigate the effect on the forecasting accuracy of GARCH models using different density functions for z t .…”